Revisiting academic integrity after GenAI: Learners’ perspectives on fundamental and novel challenges

Marian Hurley¹ , Deirdre Casey² , Diarmuid Grimes³ , Violeta Morari⁴

¹Department of Marketing and International Business, Munster Technological University, Ireland, ²Department of Mathematics, Munster Technological University, Ireland, ³Department of Computer Science, Munster Technological University, Ireland, 4Department of Mathematics, Munster Technological University, Ireland.

How to cite: Hurley, M.; Casey, D.; Grimes, D.; Morari, V. (2026). Revisiting academic integrity after GenAI: Learners’ perspectives on fundamental and novel challenges. In: 12th International Conference on Higher Education Advances (HEAd’26). Valencia, 15-18 June 2026. https://doi.org/10.4995/HEAd26.2026.21740

Abstract

This paper presents findings based on our analysis of qualitative data on students’ concepts of academic integrity collected via survey at an Irish university during the 2023-24 academic year. It retrospectively reviews the supports our respondents wished to see implemented to support them in acting with integrity in their learning and assessment practice. We find that while generative artificial intelligence featured as a prominent educational integrity concern, still more urgently felt was the need for education, information, and guidance on the meaning and form of academic integrity, as well as clarity on what academics are asking for when they stipulate that learners should enact it. We read this finding in light of the roughly contemporaneous emergence of GenAI-oriented assessment frameworks such as the Artificial Intelligence Assessment Scale or AIAS (Perkins, Furze, et al., 2024)
and argue that while such frameworks are a necessary component of academic integrity education post-GenAI, their effectiveness may depend upon their integration with or embedding in a more holistic education on what it required in order to act with educational integrity.

Keywords: academic integrity; generative artificial intelligence; academic literacies; learning, teaching and assessment

1. Introduction

Recent scholarship on Higher Education assessment has foregrounded the tension between communication-based “discursive” (Corbin et al., 2025a, p. 5)
approaches to curbing generative artificial intelligence (GenAI)-assisted misconduct and “structural” (p. 6)
assessment reform. This tension arises from the emerging consensus that GenAI has rendered misconduct detection unreliable (Ardito, 2025; Corbin et al., 2025b; Perkins, et al., 2024b)
, and that discursive approaches – such as the Artificial Intelligence Assessment Scale (AIAS) (Perkins et al., 2025; Perkins et al., 2024a)
– rely on learner compliance, which cannot be assured. This animating tension in the contemporary literature of academic integrity and ethics, we argue, belies an urgency on the part of Higher Education students better to understand the fundamentals of good assessment and learning practice.

This paper presents data collected via survey at a multi-campus Irish university during the 2023-24 academic year. We look retrospectively at qualitative data on students’ wishes in terms of academic integrity support at that time; that is, roughly contemporaneously with the publication of the first iteration of the AIAS (Perkins et al., 2024). We show that students’ thoughts on academic integrity assurance were quite complex and were not as exclusively bounded to the impact of GenAI as the recent scholarly literature has tended to be. While the data do reveal significant doubt among students regarding the renegotiated boundaries of acceptable academic behaviour in light of GenAI, the most pressing need they articulate is for education and guidance on the concept of academic integrity, about which they continue to experience ambivalence.

Such legacy concerns go largely unaddressed by models such as the AIAS and, critically, students’ calls for regular and incremental training on good learning and assessment practice risk going unheeded amid the urgency surrounding the impact of GenAI. Our work suggests that for GenAI integrity frameworks and paradigms to measurably impact learners’ integrity, they could usefully be embedded in relational and educational interventions on fundamental elements of good learning and assessment practice and taking the form of regular in-lecture information from lecturers, additional or supplementary lectures or modules, and readily accessible asynchronous integrity support materials.

2. Methods

With approval from our Human Research Ethics Committee (MTU-HREC-MR-23-024-A), we distributed a 22-item survey to a representative sample of the approximately 15,500 students in our university. Students from all faculties participated, and a total of 608 completed the survey. Of these, only 30 were postgraduates, therefore, we have elected to report only on undergraduate responses here to ensure reliability and generalisability. The aim of the survey was to investigate learners’ perceptions of and thoughts about academic integrity and misconduct at a critical point. Here, we analyse data gathered in response to the following two open-ended questions:

• In your view, what more could the university do to help students to avoid engaging in Academic Misconduct? and

• Do you have any other comments regarding Academic Integrity / Misconduct?

Reflexive thematic analysis of all valid answers to the qualitative questions was conducted by the two first authors. Codes assigned during the first phase were compared, and a high degree of intercoder agreement was noted. Minor variations were reconciled to create the final code frame, which was independently reapplied by both researchers, yielding almost identical assignment of codes. As such, the final set of codes was grouped into salient themes.

3. Literature review

Traditionally, the literature of academic integrity has tended to the assumption that the essence of integrity assurance lies in the circumscription, detection, or sanctioning of misconduct. Early studies discussed the (in)effectiveness of institution-wide “honour codes” (e.g. McCabe et al., 2002)
, but this focus is being supplanted apace by a more recent literature of assessment security (Birks & Clare, 2023; Ellis & Murdoch, 2024; Lancaster, 2023)
which, despite the relatively novel approaches it synthesises, continues to assume an interdependence between integrity and misconduct management. Fostering AI literacy is another commonly proposed corrective to the integrity threats posed by GenAI (Barrett & Pack, 2023)
, although this is problematised by the difficulty of adequately defining the concept.

Equally prominent in contemporary scholarship is the debate around assuring the validity of assessment instruments in light of GenAI. Corbin et al. (2025b)
argue that in the GenAI era, assessment processes should encompass elements which cannot feasibly be completed by GenAI. Applicability, however, may vary based on, for instance, cohort size or disciplinary constraints. Additionally, this approach, too, equates validity with the necessity of completing at least some assessment elements without access to GenAI.

The AIAS represents a holistic approach to designing and implementing assessment with GenAI. Its five levels – ranging from “no AI” to “AI exploration” (Perkins et al., 2025)
– support lecturers and students by allowing them to negotiate and define the extent to which and purposes for which GenAI may be used by the student in the completion of an assignment. Its effectiveness, however, depends on students’ willingness to accept and follow its guidance. Pilot testing does appear to validate its positive impact on the cultivation of academic integrity with GenAI integration (Furze et al., 2024)
. In spite of its evident promise, it conforms to the integrity-as-misconduct-mitigation paradigm, positing that defining a GenAI “limit” can circumvent misconduct and, therefore, promote integrity.

4. Findings

4.1. Respondents’ views on academic misconduct and sanctions

Of 52 valid responses coded under “consequences of/for academic misconduct”, 22 expressed a perceived need for more, stricter, or more consequential “punishment” for assessment infractions. Suggestions range from implementing stricter assessment rules to expulsion, which one respondent notes is “an effective deterrent.” Similar sentiments are expressed by another who is of the view that expulsion should result in the case of a second infraction.

Perhaps more consequentially, however, 27 indicate a desire for information and education, both on how to avoid academic misconduct and on the consequences of failing to do so. One respondent, then in the first year of their studies, captured both concerns:

I think [the university] should clearly lay out the consequences that students will have to face if they have participated in Academic Misconduct. By clearly laying out consequences for their actions students might be less inclined to do academic misconduct.

A third-year student highlighted the examination setting, suggesting that “before an exam there should be a clear consequence made to us [sic.] in regards to what happens if you're caught cheating.” However, frequency of communication was a more widespread concern, with five respondents urging the university to design academic integrity education such that it is incremental and regular.

Three further responses point – if indirectly – to the influence of lecturers in the assessment grading process. A first-year learner, for instance, states that they avoid GenAI use as they are “scared of repercussions.” A third-year participant favours “fewer penalties and more understanding” (emphasis added). It is also suggested that lecturers, through their monitoring and grading of assessment, are ultimately the arbiters of assessment integrity. Of this, a final-year student believes “it's unfair when you put effort in and others use chat gpt and get a good score even though we are told we will be deducted marks, but they never get deducted!”

Eight responses address the question of unfair advantage. Responses in this code, too, reflect a concern with the lecturer’s centrality to ensuring that undue advantage is avoided:

Some lecturers set time restrictions on labs to help prevent chatgpt usage. This is not acceptable because people can still do it inside of the lab and it’s not fair to other students. I myself am staying away from that stuff because I want to challenge myself and I see it as a bit of a cheap way to get good marks.

In contrast, a belief that it is incumbent upon lecturers to integrate GenAI into the learning and assessment environments is also evident:

Only one lecturer has ever acknowledged that ai such as chatgpt can be a helpful learning tool. A lot of students including myself use it for summarising notes and simplifying difficult to comprehend concepts, not to copy and paste assignments. There needs to be some acknowledgment that not all ai is bad and it doesn't always equal plagiarism, as technology and facilities advance so should teaching.

This tension is, of course, closely replicated in the literature of Higher Education assessment and remains, as yet, unresolved. It is notable, however, that our survey respondents largely want a top-down response where the “new acceptable” is defined, communicated, and monitored by the institution and by academics.

4.2. Renegotiating the boundaries of the acceptable: what and who?

4.2.1. The institution

77 responses suggest that responsibility for ensuring that learners are fully conversant with the norms of academic integrity lies with the university. This is consonant with Corbin et al.’s (2025c, p. 12)
assertion that “the absence of clear institutional guidance forces them [students] into sophisticated but unsupported forms of boundary work.” While two responses state that existing measures are adequate, 50 call for additional synchronous education within and supplementary to their lectures. Rationales include the cultivation of “[p]roper awareness of academic misconduct” as “sometimes may not know they are engaging in it.” Such sessions are also framed as an alternative to asynchronous resources which can easily be overlooked: “the guidelines we have on Academic Integrity are done in a "learn if you want to" way, and the people who look at them aren't the people who need them.”

Presumably since the parameters continue to shift rapidly, one respondent considers that the university should “make every student attend a talk monthly about the importance of avoiding misconduct” (emphasis added). Finally, related but significantly smaller codes grouped seven comments suggesting that specific education on the consequences of academic misconduct is needed and a further nine arguing that additional asynchronous resources would suffice.

Ten students invoked the need for specific education on the interface between GenAI and academic integrity. Rationales include the need for acceptable and unacceptable uses of GenAI – and of specific GenAI tools – to be made clear. While one student believes that “the use of GenAI should be barred from students. Rather I believe that students should be equipped with strategies on how to use GenAI most effectively in their area- and more importantly, equipped with the knowledge of where and when GenAI can fall short as an otherwise helpful tool”, another would like to see a “talk about what online assists can be used for example grammarly and explain how ai things don't often come out with a better answer than you do because they are often out of context.” Importantly, mirroring the above call for “unavoidable” synchronous education on academic integrity, an additional respondent notes that:

A lot of code of conduct most people just push the accept button so that they can continue with out actually knowing so I think a course or view demonstration as to what and why they should not to use chat or and ai tools before they are allowed to start the course or log in to a pc and every so often […] refreshers would be a nice feature.

A final thread relates to openness of discussion of GenAI within the university setting, e.g. “provide more information on these sites, instead of telling us to never use it show us how it can benefit our learning.”

4.2.2 Lecturers

A significant thematic thread in the data concerned the responsibilities of lecturers in ensuring that their visions of academic integrity be clarified and communicated. 26 students believed that lecturers need to support students to produce assessments with integrity. The two key areas in which additional lecturer support were called for were in instruction, interaction and classroom discussion, and in the provision or more or clearer instructions and guidelines.

“More education about academic misconduct by lecturers” is among the most requested interventions, raised 14 times. Specific suggestions include “giving precise explanations on what the assignment expects you to do [which] leaves less room for confusion/cheating” and “interact[ion] with the class such as providing worksheets and doing activities.” More generically, a need for lecturers to “properly explain how to reference and give good resource[s] for research” was a common thread. Again, the lecture or other live learning session is identified as the preferred venue for this, being preferable to “just having a pdf.” Instead, “lecturers should dedicate a class in educating students about academic misconduct.” The literature’s growing convergence around the need to move away from detection and sanctions is mirrored by one student who believed that “people are always going to do misconduct, I don't think that will ever be eliminated. However, it can be minimized by making explanations easier and more transparency within the assignments themselves.”

Notably, too, there is a clear desire for additional or improved written guidelines and briefs. This subtheme comprises 12 contributions, of which typical responses include “lecturers making assignment briefs more clear” and “provid[ing] better guidelines on assignments that are easily followed.” Lecturers are also thought to have scope to develop “more notes for modules that are easier to learn”, “more calculation examples and formulas” and “better resources on how to properly reference material that is accessible at any time on canvas”.

5. Discussion and conclusions

Of academic integrity, one succinct response to our open survey questions states that “it can be vague.” It is this vagueness – or students’ perception of it – that risks limiting the educational impact of quickly proliferating GenAI-oriented assessment integrity frameworks. While GenAI in assessment may be a “wicked problem” (Corbin et al., 2025a, p. 2)
, our analysis suggests that, left unremedied, ambiguities among students regarding fundamental principles of “intellectual honesty” (Costa & Murphy, 2025, p. 2)
could further complicate efforts to remake the norms of academic integrity post-GenAI. Furthermore, underpinning GenAI-focussed education with foundational academic literacies (Lea &Street, 1998) – whether institutional, departmental, disciplinary, or some combination of these – may represent a workable route to mitigating the ‘wickedness’ of the problem.

Undeniably, GenAI complicates students’ perception of academic integrity, and appears to have reawakened a perceived need for monitoring of assessment behaviour and sanctioning of misconduct. However, the broader norms of ethics in learning and assessment represent a more significant immediate issue for the majority of our respondents. From the perspectives of these learner-participants, it is the place of their institutions and of their lecturers to ensure that boundaries are drawn, communicated, modelled and monitored. Indeed, much of the support which students wish to see would ideally occur in the learning environment, whether as intrinsic to module delivery or as a synchronous or asynchronous addition to it, suggesting an appreciable appetite for open discussion of integrity challenges to facilitate co-negotiation of an unstable academic environment.

Institutional adoption of GenAI-focussed integrity guidelines is a positive step toward clarity and responds to a perceived need for guidance on GenAI in assessment, but they are unlikely to constitute a comprehensive response to learners’ difficulty in identifying how best to ‘be a student with integrity’. Ultimately, our analysis suggests that it behoves institutions and educators to engage learners actively and regularly in constructive dialogue about the shifting borders of appropriate practice to ensure that they can be applied with confidence and clarity. Frameworks naturally support this dialogic process, but their scaffolded transposition to learning and teaching practices are central to their achieving their intended ends.

References

Ardito, C. G. (2025). Generative AI detection in higher education assessments. New Directions for Teaching and Learning, 2025, 11-28. https://doi.org/10.1002/tl.20624

Barrett, A., & Pack, A. (2023). Not quite eye to A.I.: student and teacher perspectives on the use of generative artificial intelligence in the writing process. International Journal of Educational Technology in Higher Education, 20. https://doi.org/10.1186/s41239-023-00427-0

Birks, D., & Clare, J. (2023). Linking artificial intelligence facilitated academic misconduct to existing prevention frameworks. International Journal for Educational Integrity, 19. https://doi.org/10.1007/s40979-023-00142-3

Corbin, T., Bearman, M., Boud, D., & Dawson, P. (2025a). The wicked problem of AI and assessment. Assessment and Evaluation in Higher Education, 1-17. https://doi.org/10.1080/02602938.2025.2553340

Corbin, T., Dawson, P., & Liu, D. (2025b). Talk is cheap: why structural assessment changes are needed for a time of GenAI. Assessment and Evaluation in Higher Education, 50(7), 1087-1097.https://doi.org/10.1080/02602938.2025.2503964

Corbin, T., Dawson, P., Nicola-Richmond, K., & Partridge, H. (2025c). ‘Where’s the line? It’s an absurd line’: towards a framework for acceptable uses of AI in assessment. Assessment and Evaluation in Higher Education, 50(5), 705-717. https://doi.org/10.1080/02602938.2025.2456207

Costa, C., & Murphy, M. (2025). Generative artificial intelligence in education: (what) are we thinking? Learning, Media and Technology, 1-12. https://doi.org/10.1080/17439884.2025.2518258

Ellis, C., & Murdoch, K. (2024). The educational integrity enforcement pyramid: a new framework for challenging and responding to student cheating. Assessment and Evaluation in Higher Education, 49(7), 924–934. https://doi.org/10.1080/02602938.2024.2329167

Furze, L., Perkins, M., Roe, J., & MacVaugh, J. (2024). The AI Assessment Scale (AIAS) in action: A pilot implementation of GenAI-supported assessment. Australasian Journal of Educational Technology, 40(4), 38-55. https://doi.org/10.14742/ajet.9434

Lancaster, T. (2023). Artificial intelligence, text generation tools and ChatGPT – does digital watermarking offer a solution? International Journal for Educational Integrity, 19(10). https://doi.org/10.1007/s40979-023-00131-6

Lea, M. R., & Street, B. V. (1998). Student writing in higher education: An academic literacies approach. Studies in Higher Education, 23(2), 157-172. https://doi.org/10.1080/03075079812331380364

McCabe, D. L., Klebe Treviño, L., & Butterfield, K. D. (2002). Honor codes and other contextual influences on academic integrity: A replication and extension to modified honor code settings. Research in Higher Education, 43, 357-378. https://doi.org/10.1023/A:1014893102151

Perkins, M., Furze, L., Roe, J., & MacVaugh, J. (2024a). The Artificial Intelligence Assessment Scale (AIAS): A Framework for Ethical Integration of Generative AI in Educational Assessment. Journal of University Teaching and Learning Practice, 21(6). https://doi.org/10.53761/q3azde36

Perkins, M., Roe, J., & Furze, L. (2025). Reimagining the Artificial Intelligence Assessment Scale (AIAS): A refined framework for educational assessment. Journal of University Teaching and Learning Practice, 22(7). https://doi.org/10.53761/rrm4y757

Perkins, M., Roe, J., Postma, D., McGaughran, J., & Hickerson, D. (2024b). Detection of GPT-4 Generated Text in Higher Education: Combining Academic Judgement and Software to Identify Generative AI Tool Misuse. Journal of Academic Ethics, 22, 89–113. https://doi.org/10.1007/s10805-023-09492-6